2566.pdf

By

Mary M. Maniscalco, M.D.

A Master's Paper submitted to the faculty of The University of North Carolina at Chapel Hill

In partial fulfillment of the requirements for The degree of Master of Public Health in

The Public Health Leadership Program

Chapel Hill

2002

Over the past several years, the number of physicians practicing as

hospitalists or using hospitalists to care for their inpatients has grown. The use of hospitalists departs from a more traditional model of practice in the U.S. in which primary care doctors care for their patients in both the inpatient and outpatient settings. Hospitalists pose a special challenge for family physicians as the

majority ofhospitalists are internists or subspecialists. Historically, the discipline of family practice in part grew out of a concern for maintaining hospital

privileges, a tradition potentially threatened by the growth ofhospitalists.

Prior research has shown that practice setting and personal characteristics influence whether or not physicians practice inpatient medicine. This paper begins to look at whether or not attitudes toward hospital care influence whether or not physicians practice inpatient medicine.

Using 2001 survey data from a random sample of practicing family physicians in North Carolina, this paper describes current self-reported practice and personal characteristics of these physicians as well as examines their attitudes toward hospital care.

attitudes, the practice of obstetrics, hospital size and post-graduate year less than 25 are positively associated with providing hospital care. In multivariate logistic regression models, attitudes toward hospital care, hospital size and post-graduate year less than 25 are positively associated with providing inpatient care.

INTRODUCTION

Over the past several years, the number of physicians practicing as hospitalists or using hospitalists to care for their inpatients has grown. The term hospitalist was first used by Wachter and Goldman in 1996 although the idea of using "dedicated inpatient specialists" had been developing and growing before that time.1 The number of practicing hospitalists has grown from less than 2,000 in 1999 to 5,000 in 2002.2•3 The use ofhospitalists departs from a more traditional model of practice in the U.S. in which primary care doctors care for their patients in both the inpatient and outpatient settings.

Nationally, many primary care physicians have questions about the role hospitalists should play in the healthcare system. 1•4•5 For instance, the majority of family practitioners still care for patients in both the inpatient and outpatient settings. 6 Hospitalists thus pose a special challenge for fumily physicians as the majority ofhospitalists are internists or subspecialists. 7 Historically, the discipline of family practice in part grew out of a concern for maintaining hospital

privileges, a tradition potentially threatened by the growth ofhospitalists.4

Family physicians cite several factors that determine why they continue to practice inpatient care. In a 1995 randomized survey of members of the American Academy ofFarnily Physicians, researchers concluded that factors "associated

with individual characteristics and choices" are most strongly associated with the

practice of inpatient medicine. 8 Younger physician age, greater enjoyment of both

inpatient and outpatient medicine, fewer extra-hospital obligations and less

complex disease in the hospitalized population were independent predictors of

physicians who did inpatient care. These correlates are similar to those recently

identified for internists. 9

Limited research shows that physicians have both positive and negative

attitudes toward the hospitalist model.5•10•11 Physicians who travel more than 15

minutes to the hospital or have an established presence ofhospitalists in the

community display more positive attitudes about hospitalists. Negative attitudes

are associated with loss of practice income, valuing the practice of inpatient

I

medicine, mandatory hospitalist systems, postgraduate year greater than 25, busier

physicians and solo practice. No association was found between age and sex of

the physicians.

With what we already know about the effects of practice setting and

personal characteristics, one could ask what role physician attitudes may play in

determining whether or not physicians practice inpatient medicine. As illustrated

in Figure 1, many factors likely influence whether or not someone cares for

patients in the hospital. It is still unclear, however, which are the most influential

determinants.

In the midst of this period of transition around inpatient care it is

important to look at attitudes and these other variables to see where we are. The

characteristics offamily physicians practicing in North Carolina in 2001. We also examined the attitudes of family practitioners toward hospital care in conjunction with personal and practice setting characteristics. This study adds to prior research by its focus on family practitioners and by looking at a statewide perspective. In addition, North Carolina is somewhat unique in that it is a largely rural state with many distinct medical markets ranging from very rural to very urban.

METHODS

Sample

A mailed survey was distributed to practicing fu.mily practitioners across the state ofNorth Carolina in 2001. The North Carolina Academy of Family Physicians endorsed the survey. Names were obtained from the Sheps Center of the University ofNorth Carolina with North Carolina Board of Medicine

verification. The fmallist consisted of a random sample of 600 family

practitioners. After the original mailing, four follow-up mailings were sent to non-responders at three-week intervals. The overall adjusted response rate, after excluding surveys that were undeliverable or those sent to retired or

non-practicing physicians, was 51%.

Questionnaire development

and who receive patients from primary care physicians and return them after the hospital care is finished."

Analysis

Univariate statistics of the survey questions were compiled after the data were examined for outliers. Bivariate analysis was then performed.

In further analysis, the outcome variable, inpatient care, was defined by the question, "Do you care for medical patients in the hospital?" The response was considered to be positive and coded as a one if respondents answered yes and was coded as a zero if respondents answered no.

The questionnaire contained a series of attitude questions related to hospital care. The questions are listed in Table 2. Physicians were asked to respond strongly agree, agree, neutral, disagree or strongly disagree to these statements. For the purposes of further analysis, the responses were recoded. A positive response to an attitude question, coded as a one, (i.e. an attitude likely to favor the practice ofhospital care) was considered to be Strongly Agree/ Agree or Strongly Disagree/Disagree and depended on the question as shown in Table 2.

When the attitude questions were further reviewed, they appeared able to be categorized into several different domains, including perceived competence, doctor-patient relationship and finance. A decision was made to focus on one question from each of these three domains based on knowledge about attitudes from the literature and potential interest to the reader.

selected from the survey based on the review of the literature or because it was felt they might be important from our statewide perspective. The variables included metropolitan versus non-metropolitan practice location, solo practice versus group practice, hospital size less than or equal to 100 versus greater than 100, post-graduate year greater than or equal to 25 versus less than 25, the practice of obstetrics and gender. A decision was made to look at post-graduate year rather than age because it was felt that this measure better operationalized practice experience. The cut-off of25 years was based on the literature.

Three separate logistic regression models were then examined using inpatient care as the outcome variable. These models were predictive models generated to see what independent variables might predict providing inpatient care. One independent variable in each model was an attitude question from one of three domains previously discussed- perceived competence, doctor-patient relationship, or finance. For the initial full models, the demographic and personal characteristic variables defmed above were also used.

The Institutional Review Board of the University ofNorth Carolina School ofMedicine reviewed and approved this research study.

RESULTS

The study population is described in Table 1. Sixty-six percent of furnily practitioners provide inpatient care while 11% practice obstetrics. Eighty-three percent of respondents do not work as a hospitalist, meaning they either do not provide inpatient care or they care for their own patients in the hospital. While 16% of furnily physicians in North Carolina practice as rotating hospitalists, only 1% work as full-time hospitalists. Thirty-five percent of family physicians work

i

generally increases practice income. Thirty-eight percent strongly agree or agree that it does while 35% strongly disagree or disagree.

In Table 3, one sees the differences in attitudes toward hospital care between those family physicians who practice inpatient medicine and those who do not. There was a statistically significant difference in responses to all attitude questions between these two groups. For example, 92% of physicians who care for their own patients in the hospital believe that personal knowledge of the patient is important compared to 75% of physicians who provide outpatient care only. Physicians who provide outpatient care only are more likely to be satisfied with their current arrangements for providing hospital care (87% vs. 72%). Physicians who provide inpatient care versus physicians providing outpatient care only are much more likely to strongly agree or agree that it is valuable for patients to have their primary doctor coordinate hospital care, believe their patients expect them to provide hospital care, and believe inpatient care is an important source of continuing medical education. When one examines most differences, however, the two groups generally respond similarly to the questions but to different degrees.

statistically significant differences, rotating hospitalists are very similar in most attitudes except as related to patient expectations. Seventy-six percent of physicians who care for their own patients in the hospital strongly agree or agree that their patients expect such care, while 54% of rotating hospitalists responded in the same manner.

No statistically significant associations between providing inpatient care and practice location, practice type or gender are found in bivariate analysis. Family physicians that practice obstetrics appear to be more likely to provide inpatient care than those who do not (91% versus 65% ), however, statistical analysis was limited secondary to small sample size. Physicians who practice in smaller hospitals (of less than or equal to 100 beds) or who are less than

post-I

Some statistically significant associations between demographic/personal characteristics and key attitude questions from the three selected domains described in the methods sections are found (perceived competence, doctor-patient relationship and finance). Doctors who practice in hospitals ofless than or equal to 100 beds were more likely to strongly disagree/disagree that medical subspecialists provide better care (93% vs. 77%). Doctors in solo practice (56% vs. 34%) and doctors in non-metropolitan areas (54% vs. 29%) are more likely to strongly agree/agree that doing hospital care generally increases practice income.

medical admissions, medical subspecialists provide better care than family practitioners had 4.5 (95% CI: 2.4, 8.5) times the odds of providing inpatient care than those who did not. Family physicians who strongly agree or agree that doing hospital care improves the doctor's relationship with his/her patients had 4.1 (95% CI: 2.3, 7.4) times the odds of providing inpatient care than those who did not. Finally, physicians who strongly agree or agree that doing hospital care generally increases practice income had 2.1 (95% CI: 1.2, 3.6) times the odds of providing inpatient care than those who did not.

Pairwise correlations between predictor variables were examined. Several pairs were found to be somewhat correlated; however, the correlations are likely

'

significant interactions between independent variables.

I

to be too weak to have much of an effect on the final model. There were no

DISCUSSION

This paper describes self-reported practice and personal characteristics of a random sample offamily physicians in North Carolina in 2001. The majority of family physicians continue to practice in the hospital as well as in the outpatient setting, while 11% practice obstetrics. Nationally, approximately 85% offumily physicians have hospital privileges and 22% practice obstetrics, higher than that seen in North Carolina.6 The numbers for North Carolina in this survey may appear lower than the numbers in the national survey because they reflect what physicians are actually doing and not what they have privileges to do. It also may be that the scope of practice, including inpatient care and obstetrics, is narrower in this region of the country than in the mid-West or West. Some fumily physicians are practicing as rotating hospitalists; however, only one percent have become full-time hospitalists.

The analysis showed that practicing in a smaller hospital is positively associated with providing inpatient care, a finding not previously described in the literature. It may be that physicians practicing in smaller hospitals may also be more likely to practice in areas with fewer providers. Thus, doing hospital care becomes a necessity, as there are insufficient providers of hospital care.

The association between providing hospital care and post-graduate year less than 25 was not entirely expected. Physicians who are post-graduate year greater than 25 have more negative views toward hospitalists, yet they are less likely to do hospital care. Perhaps physicians later in their careers have other commitments, such as administrative positions, or are scaling back their work

hours, which makes them less likely to practice in the inpatient setting despite their views on hospitalists. Alternatively, it may reflect the demands of providing both inpatient and outpatient care and, over time, physicians opt out of hospital care.

The positive association between providing hospital care and selected attitudes was something we expected to see. A well-known theory from the behavioral science literature, the Theory ofReasoned Action, suggests that

attitudes are one of the main determinants of behavior. 12 Although usually applied to patients, the model has been applied to physicians in terms of physician

behavior and clinical decision-making.13 Perhaps attitudes also have an effect on decisions physicians make regarding their scope of practice. Physicians who have a more positive attitude toward hospital care may be more likely to do hospital care. The results of this study suggest an association. However, one can't show causality with cross-sectional data. It will be interesting to see if there are shifts in attitudes and behavior as residents and practicing physicians gain more exposure to the hospitalist model of care.

Alternatively, one might suggest that attitudes reflect but don't determine behavior. If a physician is in a role where they do hospital care and they are satisfied with their position, they might report positive attitudes toward hospital care. If, however, they practice outpatient medicine only and are equally satisfied with their practice arrangement, they may say that the practice of inpatient

Physicians who like inpatient medicine, for example, may be more likely to make that an important fuctor when selecting a position. Physicians may also choose practice opportunities because of input from a spouse or significant other, specific location features other than metropolitan and non-metropolitan location, or because of other favorable characteristics, such as work hours or opportunities to teach. Thus, whether or not one does hospital care may only be a small factor in the decision-making process for some.

Overall fmdings about providing inpatient care and attitudes about practice income are quite interesting. Although physicians who do inpatient care were more likely to report that hospital care generally increases practice income than those who did not do inpatient care, the majority ofboth groups reported that hospital care does not significantly increase practice income. Perhaps physicians who add on hospital care to a full outpatient schedule, before and after their office hours, or physicians who have a higher volume of inpatients may fmd greater financial benefits than others. Although enjoyment of the practice of inpatient medicine, altruism and perceptions of the importance of continuity of care are clearly important, one has to question how long these values will continue to influence family physicians to keep practicing inpatient medicine when economic constraints persuade in the opposite direction.

Some of our results are comparable to previously published studies. Similar to other researchers, this study found that positive attitudes toward hospital care were associated with providing hospital care. 8 Most fumily practitioners continue to retain a broad scope of practice and provide both

inpatient and outpatient care as they did in the mid 1990's.3 Our findings related to post-graduate year parallel previous work showing associations between age and providing hospital care; however, we did not find an association between hospital care and practice type. 3

A comparison of family physicians and general internists shows clear differences. In a study of general internists, seventeen percent of general internists provide outpatient care only compared with 34% of family practitioners in our study.9 As discussed previously, fewer family practitioners in our study provided inpatient care than reported in a national survey of family practitioners. 6

Although this survey did not specifically ask about attitudes toward hospitalists but rather hospital care, some findings in the literature about attitudes toward hospitalists are interesting in light of our study. One study on attitudes toward hospitalists published in 2000 found that solo practitioners had more negative attitudes toward hospitalists.10 Although one might expect this to make them more likely to do hospital care, we did not find this association to be true. Our results are comparable with another survey in which 54% of physicians felt hospitalists would hurt established doctor-patient relationships.11 In this study, greater than 70% of physicians felt providing hospital care improved the doctor-patient relationship.

instrument was a survey, it assumes that respondents accurately report their scope of practice. This assumption is more likely to be true as results were not

identifiable, and respondents therefore had no reason to not answer honestly. Generalizability is limited as the responses to the survey reflect the practices, characteristics and attitudes offumily physicians in North Carolina only.

However, as most focus has been on internists or family practitioners on the West Coast, this snapshot ofNorth Carolina adds diversity to the literature. Given the small number offamily physicians sampled who practice as full-time hospitalists, one can make no comment about attitudes of full-time hospitalists from this survey.

The

between those physicians who practice inpatient medicine versus those who do not has implications for the training and recruiting of future family physicians. Perhaps it is important to instill or support positive attitudes toward inpatient care during residency to encourage family physicians to actively participate in hospital care in or out of a hospitalist model. One might also want to see if family

settings. Increasing financial reimbursement may improve satisfaction but goes against current trends by both public and private payers.

One has to question whether differences in attitudes may also have an effect on the growth and use of full-time hospitalists. If a significant number of family physicians strongly favor practicing in both the inpatient and outpatient setting, will this slow down the growth ofthe full-time hospitalists?

The association between smaller hospital size and the practice of inpatient medicine suggests there is a threshold for hospital size below which a hospital cannot support the services of a hospitalist. If this is the case, one cannot lose the

l

I

both the inpatient and outpatient settings. A larger sample may also allow one to look at potential differences in scope of practice and attitudes of family physicians that practice obstetrics. Are these the subset of family practitioners one would target to maintain both inpatient and outpatient skills? A related training issue then is whether or not two tracks will develop in internal medicine and family practice residencies with one geared toward inpatient medicine and the other geared toward outpatient care.

training needs might be different as many of those issues are already emphasized in family practice residency training programs. Additionally, ifhospitalists after several years want to practice in the outpatient setting, what will their re-training needs be and how will that be accomplished?

Hospitalist services are also growing in academic medical centers and this growth may effect the current training of residents. 3 In one study, the hospitalist service was felt to provide a "better educational enviromnent" although some residents expressed concern about loss of autonomy and exposure to fewer faculty and the diversity that comes with that.15 Less is known about the effects of exposure to hospitalists on medical students. 16

In terms of patient care, a recent review of the literature shows that the use ofhospitalists leads to improved or stable quality of care and patient satisfaction.3 This is an area for further research as the number of studies and quality of the methodology is of some concern. Another area known to need more attention is improving communication between hospitalists and primary care physicians. In a recent study, only 56% of primary care physicians were satisfied with

communication. 17 The survey instrument described in this study contains valuable information about communication issues and the perceptions of fumily physicians in North Carolina and will be analyzed at a later date.

Metropolitan

vs.

non-metropolitan

practice location

~

j

Attitudes toward hospital care

j

/

Solo practice

vs.

group practice

Hospital

size-small vs. large

\

Years post-graduate

experience

I

Provides Hospital

Care

t

Provides obstetrical care

vs.

no obstetrical care

\

[ae~d~~··l

n=277 Gender:

71%Male 29% Female Age (Mean+/- SD):

45 +!- 11 years Race:

83% White

7% African American 11% Other

Practice Location:

65% Metropolitan 35% Non-metropolitan Practice Characteristics:

Size of practice:

Solo 19%

Small FP group (2- 5) 33%

Medium or Large FP group 19% FP within a multispecialty group 8%

Other 21%

Scope of practice:

Provide inpatient care 66%

Practice obstetrics 11%

Provide prenatal care 16%

Provides newborn care 51%

Provide psychological counseling 57% Manage patients in nursing homes 46% Hospital Size:

<50 Beds 8%

50-lOOBeds 21%

101-500 Beds 52%

> 500 Beds 19%

Scope ofHospitalist Practice:

Do not work as a hospitalist 83% Work as a rotating hospitalist 16%

2001

Levels of Agreement or Disagreement

Statement SA A N

"I am satisfied with my current arrangements 35% 42% 11% for the hospital care of my patients"

"For routine medical admissions, medical <1% 5% 14% subspecialists provide better care than FPs"

"Personal knowledge of a patient is important 42% 44% 12% in hospital care"

D SD

11% 1%

46% 35%

1% <1%

"It is valuable for patients to have their 27% 46% 20% 6% <1% primary doctor coordinate hospital care"

"Doing hospital care is an important source 26% 42% 20% 11% 1% of continuing medical education"

"Doing hospital care improves the doctor's 31% 46% 20% 3% <1% relationship with his/her patients"

"My patients expect me to care for them in 27% 27% 22% 15% 9% the hospital"

"Doing hospital care generally decreases time 50% 37% 6% 5% 1% for the doctor's family"

"Doing hospital care generally decreases the 37% 37% 12% 12% 2% doctor's efficiency in the office"

2001 n =277

Physicians who provide "inpatient and outpatient care" versos "outpatient care" only

Attitudes

"I am satisfied with my current arrangements for the hospital care of my patients"(% SA/A)

Provide Outpatient and Inpatient

Care

72%

"For routine medical admissions, medical 89%

subspecialists provide better care than FPs" (% SD/D)

"Personal knowledge of a patient is 92%

important in hospital care" (% SA/ A)

"It is valuable for patients to have their 83%

primary doctor coordinate hospital care" (% SA/ A)

"Doing hospital care is an important source of 80%

continuing medical education"(% SA/A)

"Doing hospital care improves the doctor's 85%

relationship with his/her patients" (% SA/ A)

"My patients expect me to care for them in 71%

the hospital" (% SA/ A)

"Doing hospital care generally decreases time 9%

for the doctor's family"(% SD/D)

"Doing hospital care generally decreases the 18%

doctor's efficiency in the office" (% SD/D)

"Doing hospital care generally increases 43%

practice income" (% SA/ A)

Provide Outpatient

Careon!y p

87% .008

64% <.001

75% <.001

54% <.001

44% <.001

58% <.001

18% <.001

0% .003

7% .013

2001

Physicians Caring for their own Inpatients versus Practicing as a Rotating Hospitalist

Care for their Practice as own inpatients a rotating

hospitalist p "For routine medical admissions, medical 90% 74% .002 subspecialists provide better care than FPs" (% SD/D)

"It is valuable for patients to have their 85% primary doctor coordinate hospital care" (% SA/ A)

"My patients expect me to care for them in the hospital" (% SA/ A)

76%

70% .007

Association between attitudes and odds of providing inpatient care for key attitude questions

Unadjusted Models

SD/D For routine medical admissions, medical subspecialists provide better care than FPs SA/A Doing hospital care improves the doctor's relationship with his/her patients

SA/ A Doing hospital care generally increases practice income

Final Predictive Models

Modell- Perceived Competence Domain SD/D For routine medical admissions, medical subspecialists provide better care than FPs Hospital Size < or = 100 beds

Post-grad year> or= 25

Mode12- Doctor-Patient Relationship Domain SA/A Doing hospital care improves the doctor's relationship with his/her patients

Hospital Size< or= 100 beds Post-grad year> or= 25 Model3 - Finance Domain

SA/ A Doing hospital care generally increases practice income

Hospital Size < or = 100 beds Post-grad year> or= 25

Odds Ratio (95% CD 4.5 (2.4, 8.5)

4.1 (2.3, 7.4)

2.1 (1.2, 3.6)

Odds Ratio (95% Cl) 4.2 (2.1, 8.5)

2.5 (1.2, 5.2) .42 (.21, .83)

5.5 (2.8, 10.9) 3.7 (1.7, 8.0) .38 (.19, .75)

1. Weissler J. The Hospitalist Movement: Caution Lights Flashing at the

Crossroads. Am J Med. 1999;107:409-413.

2. Lurie JD, Miller AD, Lindenauer PK, et al. The Potential Size of the

Hospitalist Workforce in the United States. Am J Med. 1999;106:441-445.

3. Wachter JM, Goldman L. The Hospitalist Movement 5 Years Later. JAMA.

2002;287:487-494.

4. Geyman JP. Family Practice in a Failing Health Care System: New

Opportunities To Advocate for System Reform. JABFP. 2002;15:407-416.

5. Fernandez A, Grumbach K, Goitein L, eta!. Friend or Foe? How Primary Care

Physicians Perceive Hospitalists. Arch Int Med. 2000;160:2902-2908.

6. Personal communication (Greg Tolleson- AAFP 1 0/8/02): Unpublished

Practice Profile Survey- 5/02.

7. Lindenauer PK, Pantilat SZ, Katz PP, Wachter RM. Hospitalists and the

Practice oflnpatient Medicine: Results of a Survey of the National Association

Current Inpatient Practice. JABFP. 1997;10:357-362.

9. SaintS, Konrad T, Golin C, et al. Characteristics of General Internists Who Practice Only Outpatient Medicine: Results from the Physician Worklife Study.

Sem Med Pract. 2002;5 :5-11.

10. Auerbach AD, Nelson EA, Lindenauer, et al. Physician Attitudes toward and Prevalence of the Hospitalist Model of Care: Results of a National Survey. Am J

Med 2000;109:648-653.

11. Auerbach AD, Davis RB, Phillips RS. Physician Views on Caring for Hospitalized Patients and the Hospitalist Model oflnpatient Care. JG/M

2001;16:116-119.

12. Fishbein M. A Theory ofReasoned Action: Some Applications and Implications. Nebraska Symposium on Motivation. 1979:65-116.

13. Millstein SG. Utility of the Theories of Reasoned Action and Planned Behavior for Predicting Physician Behavior: A Prospective Analysis. Health

Residency Training Needs: Results of a National Survey. Am J Med.

2001;111 :247-254.

15. Chung P, Morrison J, Jin L, eta!. Resident Satisfaction on an Academic Hospitalist Service: Time to Teach. Am J Med. 2002;112:597-601.

16. Hauer KE, Wachter RM. Implication of the Hospitalist Model for Medical Students' Education. Academic Medicine. 2001;76:324-330.